54 un J C/in Niar 1993:58:54-60. Printed in USA. � 1993 American Society for Clinical Nutrition

Delayed lactogenesis in women with insulin-dependentdiabetes mellitus14

Su:anne H Neuhauer. .4 nn lvi Ferris. Christina G Chase, Joanne Fane/li, Co/leen A Thompson.

Carol J Lammi-Keefe. Richard M Oark. Robert G Jensen. Robert B Bendel. and Karen W Green

ABSTRACT Breast milk lactose. total nitrogen. conductiv-

ity. osmolality. and intake by infants of33 women with insulin-

dependent diabetes mellitus (IDDM). 33 control women without

diabetes. and I 1 reference women were determined in a 3-mo

study oflactation. Milk ofwomen with IDDM had significantly

lower lactose and higher total nitrogen (2-3 d postpartum). and

their infants had significantly less milk intake (7-14 d postpar-

tum) than did control or reference women. Total nitrogen was

negatively correlated with milk lactose for women with IDDM

at all times and for control women through day 14 postpartum.

The data indicate delayed lactogenesis for women with IDDM.

which was more likely to occur with poor metabolic control.

Differences in milk composition ofwomen with IDDM do not

preclude them from breast-feeding their infants. .4,ii J Clin

Nutr 1993:58:54-60.

KEY WORDS Lactation. initiation of lactation. breast-

feeding. human milk. milk composition. insulin-dependent di-

ahetes mellitus, metabolic control. breast milk intake. test

weighing

Introduction

The ability ofsome women with insulin-dependent diabetes

mellitus (IDDM) to breast-feed has been documented ( 1 . 2) as

has preliminary information on the unique aspects oftheir din-

ical management (2). Although women with IDDM choose to

breast-feed as often as do women without diabetes (2). clinical

research concerning lactation in women with IDDM is limited.

Milk lactose is thought to be a marker for lactogenesis (3. 4).

Lactose increases while protein. sodium. and chloride decrease

(5. 6): these changes precede the increase in milk volume secre-

tion (7). Arthur et al (4) found that the onset ofcopious milk

secretion at 24-48 h postpartum, was delayed significantly in

women with IDDM: however, breast milk lactose reached a pla-

teau by day 4 postpartum in women with and without IDDM.

Bitman et al (8) studied one woman with IDDM and determined

that milk volume at 3 d postpartum was lower than that of a

normal reference group but normalized by day 5 postpartum.

Euglycemia may influence milk composition and ultimately

affect lactation outcome. Hyperglycemia in alloxan-induced

diabetic rats or goats resulted in decreased milk lactose and

protein (9. 10). Milk yields were also decreased. Insulin ad-

ministration reversed these effects: however, permanent loss

of mammary gland function occurred when insulin was with-

held ( I 1 ). At 90 d postpartum. Butte et al ( 1 2) studied breast

milk ofwomen with IDDM who had elevated hemoglobin (Hb)

A,� concentrations. Breast milk lactose and protein were the

same as that ofa reference population but glucose and sodium

were significantly higher. Hypoglycemia may also occur in

women with IDDM: it has been found in the early postpartum

period and after breast-feeding (2. 13). Hypoglycemia in ani-

mals increased epinephrine. which in turn decreased blood flow

to the mammary gland ( I 4. 1 5). Decreased lactate with in-

creased ketone production was accompanied by decreased

mammary gland lipogenesis (16). Ultimately breast milk corn-

position may be affected.

Because breast milk composition ofwomen with IDDM and

intake by their infants have not been evaluated longitudinally,

the present study investigated lactation in women with IDDM

by determining for 3 mo breast milk composition of lactose,

total nitrogen. conductivity. osmolality, and intake by infants

of women with IDDM. Electrical conductivity was used to

measure anion and cation changes in milk that occur with the

initiation oflactation ( I 7). Ifmilk ofwomen with IDDM differs

in lactose and electrolytes. then milk osmolality might also be

affected. Morriss ( 1 8) has reviewed the effects of hyperosmolar

milk on infant conditions. Because women with IDDM must

depend on exogenous insulin to maintain euglycernia and

fine-tuned control is difficult to achieve, we also wished to

study the effect of metabolic control in women with IDDM

on the above markers of lactogenesis and intake by their

infants.

I From the Department of Nutritional Sciences and the Center forEnvironmental Health. Department of Animal Science, University ofConnecticut. Storrs. and the Division of Maternal-Fetal Medicine. Uni-

versity ofMassachusetts Medical School. MCCM-Memorial. Worcester.

MA.2 Scientific Contribution Number 1385. Storrs Agricultural Experiment

Station. University ofConnecticut. Storrs.3 Supported in part by funds made available through NICHHD ROl-

HD20208. the Hatch Act. the University of Connecticut Research

Foundation. and the Pediatric Nutrition Dietetic Practice Group of theAmerican Dietetic Association.

4 Address reprint requests to AM Ferris. Department of Nutritional

Sciences. U- I 7, 3624 Horsebarn Road Extension. University of

Connecticut. Storrs. CT 06269-40 17.Received February 25. 1992.Accepted for publication January 19. 1993.

at HO

SP

ITA

L GR

AL U

NIV

ER

SIT

AR

IO A

LICA

NT

E B

IBLIO

TE

CA

on July 18, 2014ajcn.nutrition.org

Dow

nloaded from

DELAYED LACTOGENESIS AND DIABETES 55

Methods

Subjects

The study population consisted of 33 lactating women with

IDDM, 33 lactating women without IDDM (control subjects),

and 1 1 healthy reference women as outlined previously (19).

The study had human subjects approval by all hospitals used

for recruiting (19) and by the University ofConnecticut. Women

were recruited through private obstetricians’ offices and high-

risk prenatal screening clinics at three medical centers. Data on

breast milk intake were obtained from a subgroup of mothers

who did not differ from the study population with respect to

subject characteristics. This subgroup was determined solely by

the willingness of the mother to participate. This subgroup in-

eluded 18 lactating women (control) pair-matched to 1 8 women

with IDDM. 10 reference women, and their infants. Matching

characteristics of the subgroup were as follows: gestational age.

< 37 (n = 1) and � 37 wk (n = 1 7): delivery method. vaginal

(n = 7) and cesarean section (n = 1 1 ): infant sex, males (n = 12)

and females (n = 6): prior lactation, yes (n = 9) and no (n = 9).

Five ofthe 10 reference women had prior lactation experience.

and the sex of their infants was 5 males and 5 females.

(‘ollection procedures

Subjects were visited at days 2 and 3 postpartum in the hospital

and at days 7 ± 1, 14 ± 2. and 42 and 84 ± 4 postpartum in

the hospital or at home ( 19). Blood samples were drawn 80 mm

after breakfast. HbA1� samples. drawn at 3 and 42 d postpartum,

were transported immediately on wet ice packs to the hospital

laboratory or sent by overnight mail in styrofoam containers.

HbA� samples were sent to Smith Kline Bio-Science Labora-

tories (63.0%) (Waltham, MA) or Nichols Institute (37.0%) (San

Juan Capistrano. CA) for analysis. An HPLC analyzer and Hb

autodiluter semiautomated system were used. There were no

significant differences in average HbA1� concentrations with re-

spect to the laboratory where analyzed: therefore, the data were

combined for subsequent analyses. Capillary blood samples were

analyzed for glucose at each visit. Either an Accu-Chek bG

(Boehringer-Mannheim, Indianapolis) or a Glucoscan 2000

(Lifescan, Mountain View, CA) reflectance meter was used.

Both instruments were reported to accurately reflect venous

glucose (20).

All milk samples were collected by trained researchers using

an Egnell electric pump (Egnell Inc. Cary, IL). Milk expression

was done immediately after blood glucose was determined. To

ensure a representative milk sample, mothers were asked to

breast-feed their infants before breakfast, thus ensuring a I-h

interval between breast-feeding and sampling. The breast pump

was held to the breast until milk no longer came out in an even

intermittent stream, �8 mm/breast. At days 2 and 3 postpartum

the mothers’ breasts were pumped for a maximum of 5 mm if

no sample was available. The procedure always began on the

right breast and was repeated for the left breast. All analyses

were done on samples from both left and right breasts. Samples

for nutritional analyses were immediately frozen at the collection

site on dry ice and transferred to a freezer (-70 #{176}C)until ana-

lyzed. When there was adequate milk, fresh milk for conductivity

and osmolality was sampled, transported to the laboratory on a

wet ice pack, and analyzed within 4 h.

Breast milk intakes were determined at 7, 14, 42, and 84 d

postpartum by test weighing the infant before and after each

feeding for 24 h plus one more feeding. A model 3862MP8

(Sartorius, Gottingen, Germany) integrating electronic balance

equipped with a printer and infant seat was used. The balance

integrates repetitive weighings every 2 s: precision stated by the

manufacturer is ±0.05 g. The balance was placed on a level

surface and calibrated in the home by the researcher. Verbal

and written instructions were given and mothers demonstrated

their ability to use the balance. Mothers were asked to follow

their usual daily feeding routine noting any feedings in which

formula was used. For 7 ofthe 17 formula feedings, test weighing

was not possible and intake was measured as volume removed

from the bottle.

The 24-h milk intake was calculated by 1) summing the dif-

ference of the before and after weights for each feeding, 2) sub-tracting the gram intake of the last feeding from the day’s total

(24 h plus 1 feeding) because intake from the next-to-last feed

satiated the infant until the time ofthe last feed, and 3) extrap-

olating to an exact 24-h period. The data used represented time

spans of24 ± 3 h.

Aletaholic control of WOflWfl ti’ith IDDIti

Postpartum metabolic control was determined by using

White’s classification ( I 9, 2 1). HbA1� at 3 and 42 d postpartum,

and fasting blood glucose (FBG) and blood glucose at 80 mm

after breakfast at each visit. Appropriate FBG concentrations

are 3.9-6.7 mmol/L (22). Blood glucose variability can be mea-

sured through continuous monitoring ofblood glucose patterns

and calculation of the mean amplitude of glycemic excursions

(MAGE) (23). Blood glucose at 80 mm after breakfast correlates

best with MAGE (24). Determination ofMAGE was not possible

in our study: therefore, the 80-mm postprandial glucose (PPG)

value was used. PPG should be < 8.9-10.0 mmol/L at 90 mm

(22). We considered an appropriate concentration for PPG at

80 mm to be 8.9 mmol/L.

Milk sample anali’.sis

Glucose and lactose were determined by using the model 27

industrial analyzer (Yellow Springs Instrument Co. Inc. Yellow

Springs, OH) (25). Total nitrogen was determined by the micro-

Kjeldahl method (26) by using the Tecator 1009 Digestion Sys-

tem and Tecator-Kjeltec System 1002 Distilling Unit (Hogan#{228}s,

Sweden). Conductivity was determined by using a model 35

Conductance Meter (Yellow Springs Instrument Co. Inc) cali-

brated with potassium chloride (27). Osmolality was determined

by freezing-point measurement with a model 5004 �zOsmette

(Precision Systems. Inc. Natick, MA) calibrated with sodium

chloride (28). All milk samples were analyzed in duplicate.

Statistical anali’.sis

All statistical tests were done by using SAS (29, 30). When P

equalled 0.01 it is stated, otherwise P < 0.05 is assumed. Paired

t tests were used to determine differences between the left and

right breasts. Because breast milk composition did not differ

significantly between breasts, the average was calculated. A re-

peated-measures (split-plot) analysis ofvariance (ANOVA), using

the main effects of group. individual within group. time, and

group-by-time interaction, was used to determine group and

time differences for milk composition and intake by infants.

The protected least-significant difference test was used to make

contrasts between each pair ofgroup means (3 1 ). The assumption

ofcompound symmetry was checked by using the REPEA TED

at HO

SP

ITA

L GR

AL U

NIV

ER

SIT

AR

IO A

LICA

NT

E B

IBLIO

TE

CA

on July 18, 2014ajcn.nutrition.org

Dow

nloaded from

56 NEUBAUER ET AL

statement. Ifthe compound symmetry assumption was violated

(P < 0.05). the Huynh-Feldt conservative test was used to

check for interaction between group and time. For the co-

variates 80-mm PPG and time from last breast-feeding, a re-

peated-measures analysis ofcovariance was accomplished by

using S..1S PRO(’ GUt!.The approximate F statistic was used to test the significance

of group-by-time interactions for breast milk intake. The ap-

proximate F statistic is calculated from a weighted average of

the between- and within-subject mean square error terms in the

ANOVA (3 1 ). For one-sided tests a power of 80. an effect size

of 1 .2, and a significance level of0.05 were used (32). Pilot-study

data were used in the power calculations. Minimum sample sizes

for a power of 80 were exceeded. Spearman rank correlation

coefficients were used for correlations with White’s classification.

Pearson correlation coefficients were used for milk composition

and intake data.

Results

The groups did not differ concerning the time elapsed from

breakfast to the sampling of blood and milk. The women with

IDDM had a significantly longer time interval between the

mother’s last breast-feeding and when the milk sample was taken

compared with the control and reference women [least-squares

means (LSM) ± SEM: 3.76 ± 0.2, 2.56 ± 0.3, and 1.99 ± 0.3

h. respectively: F = 6. 1 5; df = 2, 73; P < 0.0 1 ]. All groups had

milk samples taken a minimum of 2 h after the last breast-

feeding. When the time interval from the last breast-feeding was

used as a covariate, it did not affect the milk-composition results.

The milk ofwomen with IDDM had significantly less lactose

(LSM ± SEM collapsed over time: 16 1 .80 ± 3.66 mmol/L) than

the milk ofcontrol (174.87 ± 2.97 mmol/L) or reference (181.12

± 3.92 mmol/L) women (F = 7.00: df = 2, 70; P < 0.0 1 ) (TableI). Milk lactose increased significantly over time to 42 d post-

partum (F = 76.22: df = 2. 70; P < 0.00 1). Breast milk lactose

was negatively and significantly correlated with breast milk con-

ductivity at 3 d postpartum for control and reference women (r

= -0.99). By day 7 postpartum all groups had negative and

significant correlations.

Women with IDDM had significantly higher milk total nitro-

gen at days 2 and 3 postpartum than did control or reference

women (F = 3.18: df 2, 68: P < 0.01). Total nitrogen was

negatively correlated (P � 0.05) with milk lactose for women

with IDDM at days 3-84 postpartum, for control women through

day 14 postpartum. and for reference women only at day 2 post-

TABLE I

Breast milk composition*

Days postpartum

Component and

group 2 3 7 14 42 84

Lactose (mmol/L)

IDDMI 95.87 ± 7.13 161 59.87 ± 4.18 1171 163.34 ± 3.02 1291 178.01 ± 3.44 1241 185.58 ± 3.83 1201 188.16 ± 4.05 118]

(‘ontrol 130.81 ± 4.58 1141 163.01 ± 3.36 124] 175.84 ± 2.97 129] I85.40 ± 3.38 [24] 92.49 ± 3.56 122] 201.65 ± 4.23 116]

Reference 142.80 ± 6.70 161 66.38 ± 4.74 II 1] 82.08 ± 4.74 11 I] 187.89 ± 4.74 II 1] 202.00 ± 4.74 [1 I] 205.60 ± 4.74 III]

Total nitrogen (gIL)

IDDM 7.78 ± ((.47 (3]� 4.78 ± 0.23 1I3I� 3.09 ± 0.15 125] 2.59 ± 0.16 124] 2.24 ± 0.18 120] 2.06 ± 0.19 118]

Control 5.46 ± ((.29 18] 3.49 ± 0.17 1221 3.02 ± 0.15 126] 2.49 ± 0.16 124] 2.16 ± 0.17 121] .89 ± 0.21 114]

Reference 5.41 ± 0.3I [61 3.38 ± 0.25 [9] 2.94 ± 0.22 [I I] 2.49 ± 0.22 II I] 2.10 ± 0.22 (I I] 1.82 ± 0.22 [11]

Conductivity (t!(t

IDDM 0.28 ± 0.02 15] 0.31 ± 0.01 [17] 0.29 ± 0.01 [15] 0.25 ± 0.01 114] 0.25 ± 0.02 [11]

Control 0.32 ± 0.02 17] 0.29 ± 0.01 120] 0.25 ± 0.01 119] 0.20 ± 0.01 119] 0.21 ± 0.01 [16]

Reference 0.34 ± 0.02 18] 0.27 ± 0.01 II 1] 0.25 ± 0.01 [I II 0.21 ± 0.01 [10] 0.20 ± 0.01 [11]

Osmolality

(mOsmoIIkg)�

IDDM 276.24 ± 9.09 [5] 297.98 ± 4.34 II6] 292.09 ± 4.55 1151 279.02 ± 4.86 [14] 286.89 ± 5.54 [II]

Control 287.16 ± 7.04 [7J 290.19 ± 3.92 [201 286.64 ± 4.22 118] 290.75 ± 3.96 [20] 298.09 ± 4.52 [16]

Reference 293.15 ± 6.01 18] 295.77 ± 4.95 [1 I] 296.95 ± 4.95 [I I] 292.79 ± 5.27 [10] 290.45 ± 4.95 [11]

Milk intake of

infants(gld)Il

1DDM� 309.62 ± 32.37 [15] 426.05 ± 32.28 115] 575.29 ± 33.98 [14] 530.93 ± 39.47 [11]

Controi� 455.46 ± 29.71 [17] 504.80 ± 35.69 1131 535.35 ± 36.04 [13] 51 1.95 ± 39.81 [II]

Reference 518.50 ± 38.20 [101 592.l7 ± 40.93 [9] 654.43 ± 38.20 [10] 673.72 ± 38.20 [10]

Milk and Ibrmula

intake of

infants (gJd(I�

IDDM 329.89 ± 27.63 I15]� 447.10 ± 27.55 [1S]** 624.70 ± 29.01 [14] 580.72 ± 33.69 [11]

Control 456.52 ± 25.36 [171 516.49 ± 30.46 [13] 542.44 ± 30.76 [13] 588.59 ± 33.98 [11]

Reference 528.12 ± 32.60 [101 595.36 ± 34.93 [9] 654.43 ± 32.60 [10] 684.33 ± 32.60110]

S Least squares means ± SEM. When group names are footnoted. group least-squares means collapsed over time are statistically different. ti in brackets.

ti: Significantly different from both control and reference: (P < 0.01, �P < 0.05.

§ Data not available at 2 d postpartum.

II Data not available at 2 or 3 d postpartum.1** Significantly different from reference: ‘P < 0.01. �P < 0.05.

at HO

SP

ITA

L GR

AL U

NIV

ER

SIT

AR

IO A

LICA

NT

E B

IBLIO

TE

CA

on July 18, 2014ajcn.nutrition.org

Dow

nloaded from

IDDM

8

6

Control

-J

0

EE

a)(I,

0

00

-J

-J

0

E

E

a)(I)0

00

-J

-J

0

EE

a)C,)0

00

-J

210

180

150

120

90

210

180

150

120

90

210

180

150

120

90

point.

Reference

,f:� I

2 3

DELAYED LACTOGENESIS AND DIABETES 57

partum. Correlations ranged from -0.57 to -0.76 for women

with IDDM and from -0.43 to -0.80 for control women; the

value for reference women was -0.87 at day 2 postpartum. We

graphed lactose and total nitrogen together to study their rela-

tionship as an index of lactogenesis. Although lactose and total

nitrogen intersect at similar concentrations for all groups, the

intersection occurs earliest in reference women, followed by

control. then women with IDDM (Fig I).

Breast-milk conductivity and osmolality did not differ among

the three groups. Milk conductivity decreased significantly over

time from 7 to 42 d postpartum (F = 21.12; df = 2, 54; P

< 0.00 1 ). Milk lactose (r = 0.36, 0.29), total nitrogen (r = -0.26,

-0.14) and conductivity (r = -0.37, -0.19) were significantly

correlated with feeding frequency for women with IDDM and

control women, respectively. Milk intake was significantly cor-

related with feeding frequency for women with IDDM at 14 (r

= 0.57) and 42 d postpartum (r = 0.8 1). for control women at

Days Postpartum

FIG I . Breast milk lactose (D) and total nitrogen (U), concentrations

for women with insulin-dependent diabetes mellitus (IDDM) and controland reference women. Least squares means ± SEM, except when theSEM is smaller than the diameter ofthe symbol. See Table 1 for number

of subjects. The point of intersection is used as an index of lactogenesis.

Lactogenesis occurred first in reference women, then in control women,

and finally in women with IDDM.

84 d postpartum (r = 0.72). and for reference women at 42 d

postpartum (r = -0.78).

Infants ofwomen with IDDM and control women consumed

significantly less breast milk (g/d) than did infants of reference

women (F = 4.69; df = 2, 43) (Table 1). Infant milk intake was

significantly correlated with breast milk lactose and conductivity

(negatively) for women with IDDM at 7 (r = 0.57, -0.81, re-

spectively) and 14 (r = 0.56, -0.76, respectively) d postpartum,

and at 7 (r = 0.68. -0.75, respectively) d postpartum for control

women. Milk intake of infants of women with IDDM was also

significantly negatively correlated with total nitrogen at 7 (r

= -0.64) and 14 (r = -0.53) d postpartum. Infants of women

with IDDM consumed significantly less breast milk and formula

combined than did infants ofcontrol women at 7 d postpartum

and of reference women at 7 and 14 d postpartum (F = 2.58;

df= 2, 43).The mean HbA1� concentration at 3 and 42 d postpartum

was within normal limits for women with IDDM: however, it

was significantly higher than in the other groups (LSM ± SEM:

IDDM 5.54 ± 0.1 1%, control 4.16 ± 0.10%, and reference 4.33

±0.l3%;F�= l4.17:df= 2, 71;P<0.00l). The majority(53%)

of the women with IDDM had an FBG � 6.7 mmol/L at all

� times. Women with IDDM had significantly higher 80-mm PPG

� concentrations than did control or reference women (LSM

� ± SEM: 13.6 ± 0.3. 5.6 ± 0.3, and 5.7 ± 0.4 mmol/L, respec-

..:. tively: F = 82.12; df = 2. 74; P < 0.001). The average PPG

4 �Q, concentration for women with IDDM ranged between I 1.6

-:: ± 0.7 and 1 5.6 ± 0.7 mmol/L throughout the study. A minimum

2 of 72% of these women had values > 8.9 mmol/L at each time

White’s classification was significantly correlated with total

8 milk nitrogen at day 3 postpartum (r = 0.38) and negatively� with lactose at days 7 (r = -0.25) and 84 postpartum (r = -0.42).

6 � There were no significant correlations between HbA� and milk

L� composition. FBG correlated significantly with total nitrogen at

:� 2 d postpartum (r = 0.99). Women with IDDM with an FBG

4 �.‘ > 6.7 mmol/L had greater milk conductivity at 7 d postpartum

> than did women with IDDM with an FBG � 6.7 mmol/L; they

/ 2 also had greater milk conductivity than control or reference

women(F= 3.17;df= 3,39:P<0.00l).

Women with IDDM with a PPG > 8.9 mmol/L had total

milk nitrogen at 2 d postpartum and osmolality at 3 d postpartum

8 significantly higher than values for women with IDDM with a

�: PPG � 8.9 mmol/L and control and reference women. Milk

6 o composition ofwomen with a PPG � 8.9 mmol/L did not differ

� significantly from that of control or reference women. Infants

4 �- ofwomen with PPG > 8.9 mmol/L had significantly less milk

.�- intake than did infants ofwomen with a PPG � 8.9 mmol/L atC � (F = 5.2 1: df = 3, 38: P < 0.0 1 ) and 14 (F = 3. 1 8: df = 3, 33)

7f’ 2 d postpartum. Milk lactose ofwomen with an FBG > 6.7 mmol/

7 L or a PPG > 8.9 mmol/L was consistently lower at 2. 7, and

14 d postpartum (NS).

When PPG was used as a covariate, infants of women with

IDDM had significantly less milk intake than did the reference

infant group at 7, 14, and 84 d postpartum and than did the

control infant group at 84 d postpartum (F = 2.27; df = 2134).

Without the covariate, group means collapsed over time were

significantly different. The adjusted group means for milk intake

for women with IDDM increased at 7, 42, and 84 d postpartum

but remained essentially unchanged for the other groups.

at HO

SP

ITA

L GR

AL U

NIV

ER

SIT

AR

IO A

LICA

NT

E B

IBLIO

TE

CA

on July 18, 2014ajcn.nutrition.org

Dow

nloaded from

58 NEUBAUER ET AL

Discussion

A delay in the initiation of lactation in women with IDDM

is indicated by the lower milk lactose and lower breast milk

intakes oftheir infants. Arthur et al (4) also found that mothers

with IDDM had a delay of 28 h for milk lactose to reach com-

parable concentrations found in women without diabetes.

The higher nitrogen content in the milk ofwomen with IDDM

at 2-3 d postpartum in the present study also suggests delayed

lactation. Kulski and Hartmann (5) found a dramatic decrease

in milk total protein consistent with the onset of lactogenesis.

To compare our results with theirs. we converted total protein

to total nitrogen. using a factor of 6.25 assuming that protein

nitrogen comprises 87% oftotal nitrogen at days 2-3 postpartum

(33). In Kulski and Hartmann’s (5) study. the derived milk total

nitrogen peaked at I 1.4 g/L 30 h after delivery and decreased

by more than halfto 4. 1 g/L by 50 h postpartum. If lactogenesis

were “on time,” the total nitrogen for the women with IDDM

(7.8 g/L) should be far less at 45 h from delivery. The total

nitrogen ofcontrol and reference women followed a similar pat-

tern of decline as that found by Kulski and Hartmann. Breast

milk nitrogen ofwomen with IDDM was not similar to that of

the other groups until 7 d postpartum.

The negative correlation between lactose and total nitrogen

for all groups at 2 d postpartum is consistent with lactogenesis.

Others have documented similar negative correlations of lactose

and whey proteins (5) or total protein (34) as estimates of the

timing oflactogenesis. In the present study the lactose and total

nitrogen contents for the reference women were correlated only

at day 2 postpartum. indicating subsequent establishment of

lactation and corroborating the findings reported previously (5.

34). The negative correlation existed the longest for women with

IDDM. indicating a delay in lactogenesis.

Saint et al (3) also found a positive correlation between lactose

and milk yield and a negative correlation between total protein

and milk yield to be consistent with lactogenesis. The lack of

significant correlations between lactose or total nitrogen and

breast milk intake for reference women further suggests that

lactogenesis likely occurred before the time when breast milk

intake was first studied at das 7 postpartum. In contrast, signif-

icant correlations still existed for control women at 7 d post-

partum and for women with IDDM at 14 d postpartum. which

suggests that full lactation could not have been established until

after this time.

Before lactogenesis. the paracellular pathway is leaky and so-

dium, chloride. and water pass into the milk diluting lactose

and maintaining osmolality (35). This additional sodium and

chloride would increase milk conductivity. However, as lactose

increases with established lactation and the paracellular pathway

no longer leaks. sodium and chloride, and consequently con-

ductivity. decrease. Delayed lactation for women with IDDM

based on milk conductivity is suggested because a negative cor-

relation with lactose was not established until day 7 postpartum

and with intake by infants until day 14 postpartum.

The lack ofearly maternal attachment associated with cesarean

section deliveries may be a factor in the delayed lactogenesis for

women with IDDM (36). Kulski et al (37) found no difference

in milk lactose as a result ofcesarean section delivery but suckling

began within 12 h after delivery in their study. Women with

IDDM first breast-fed their infants at 26. 1 ± 2.8 h after delivery

( 19): therefore. lack of early breast stimulation may have influ-

enced this delay.

It is interesting that feeding frequency was not correlated with

milk composition or milk intake for reference women but

was for women with IDDM and control women. Similarly,

Nommsen et al (38), in their longitudinal study of healthy, cx-

elusively breast-feeding women (< 120 mL/d of other milk or

formula). found that breast-feeding frequency was not significant

in the regression equations relating to macronutrient milk com-

position at 3. 6, 9. or 12 mo postpartum when controlled for

milk production. Others have also noted no relationship between

breast-feeding frequency and milk intake (39-41). It may be

that feeding frequency is not an issue for healthy breast-feeding

mothers with well-established milk output. However, for women

with multiple problems related to their prenatal care and puer-

perium recovery, such as the women in our study with IDDM

and control women, feeding frequency may be an important

determinant of breast-feeding success.

A delay in lactogenesis in women with IDDM is postulated

on the basis ofa comparison with the milk-composition values

of the reference women. Milk lactose values for the reference

women were similar to those reported at 48-72 h postpartum

(3. 7), 42 d postpartum (42), and 3 mo postpartum (38. 43).

They were slightly higher than those previously reported for milk

from our laboratory: 173 mmol/L at 42 d postpartum and 188

mmol/L at 84 d postpartum (44). Different analytical methods

may explain this because in the earlier study an enzymatic

method from Boehringer-Mannheim Biochemicals was used

(44). Our reference women had slightly lower milk lactose con-

centrations at 7 and 14 d postpartum than did multiparous

women (45) and women in developing nations (46, 47).

The milk total nitrogen concentrations ofcontrol and reference

women agreed with results ofother investigators at 1-3 (33, 48),

7-20 (44. 48. 49), and 90-105 d postpartum (44. 47-49).

L#{246}nnerdal et al reported slightly higher total nitrogen in privi-

leged women in Ethiopia (47) and Sweden (50) at 0.5-1.5 mo;

however, exact comparison is difficult because different time

periods were reported. The breast milk lactose and total nitrogen

concentrations of women with IDDM were similar to values

reported in the literature (12. 51).

The remarkable stability ofmilk osmolality found in the pres-

ent study confirms the findings ofsome researchers (28, 52, 53),

but slightly lower (54) or higher (55) osmolality was found by

other investigators. As the secretion of lactose increases to

amounts consistent with established lactation, the ionic concen-

tration decreases to maintain milk isoosmotic with plasma (35).

The similarity among the groups in milk osmolality and con-

ductivity suggests that even though women with IDDM had

significantly lower milk lactose, the ionic concentration was not

significantly altered.

Milk-intake data for infants ofwomen with IDDM from other

studies is not available. Breast-milk intake ofour reference group

was o�50 g/d below that reported by Neville et al (39) for the

same infant ages in exclusive breast-feeders. All oftheir subjects

and 50% of our subjects were multiparous. If the multiparous

women breast-fed previously, then prior lactation experience

could account for the larger intakes (56). Several investigators

have reported data at 84-90 d postpartum. Our 84-d value of

674 ± 38 g/d falls at the lower limit of values reported: 676 to

81 1 ± 133 g/d (38, 43. 57). Differences can be attributed to the

large variation in breast milk intake (43), use of a less-accurate

at HO

SP

ITA

L GR

AL U

NIV

ER

SIT

AR

IO A

LICA

NT

E B

IBLIO

TE

CA

on July 18, 2014ajcn.nutrition.org

Dow

nloaded from

DELAYED LACTOGENESIS AND DIABETES 59

1 . Whichelow MI. Doddridge MC. Lactation in diabetic women. Br

Med I 1983:287:649-50.

pediatric balance (43). or insensible water loss during breast-

feeding (38).

Although infants of women with IDDM received more for-

mula throughout the study than did the infants ofcontrol and

reference women, group differences in breast milk intake are

only offset by the higher formula intake at 42 and 84 d post-

partum. One cannot determine whether the infants of women

with IDDM who were given formula actually needed a supple-

ment. Perhaps if their mothers had not offered formula. the

infants’ demand would have stimulated increased breast milk

production to a degree that would have satisfied their appetites.

Alternatively. without the formula. their nutrition might have

been compromised.

The issue of metabolic control during lactation in women

with IDDM is not clearly defined. All criteria used to measure

metabolic control suggested that good metabolic control is an

important determinant ofsuccessful lactation, in terms of “nor-

mal” milk composition. in women with IDDM. However, no

one criterion is a predictor ofchanges in milk composition.

Women with IDDM with a higher White’s classification, and

therefore more severe diabetes and/or poorer metabolic control

(as measured by HbA1�, FBG. or 80-mm PPG). were more likely

to have delayed lactogenesis. This is indicated by the negative

correlations between White’s classification and lactose and the

positive correlations between both White’s classification and

FBG and total nitrogen at 2-3 d postpartum. As initially hy-

pothesized. women with IDDM with good metabolic control

had milk composition similar to that of control and reference

women, with respect to the nutrients measured in this study.

However, women with elevated FBG or 80-mm PPG had higher

total nitrogen at day 2 postpartum: osmolality at day 3 post-

partum: conductivity at day 7 postpartum: lower lactose at days

2. 7. and 14 postpartum: and lower infant milk intake at days

7. 14. and 84 postpartum. which suggests that the delay in lac-

togenesis was influenced by metabolic control. The increase in

the breast-milk intake of infants of mothers with IDDM when

the groups were equalized for 80-mm PPG further substantiates

this conclusion.

Women with IDDM were able to breast-feed and despite sig-

nificant differences in breast milk lactose and total nitrogen con-

centrations, milk composition for these nutrients was within

accepted ranges. There was a delay in lactogenesis for women

with IDDM, indicated by lower milk lactose. higher total nitrogen

at 2-3 d postpartum. delayed intersection oflactose and nitrogen

when graphed (Table 1 ). a negative correlation between lactose

and nitrogen. a negative correlation between both conductivity

and total nitrogen and intake by infants. and lower milk intake

by infants ofwomen with IDDM. This delay was correlated with

poor metabolic control. 13

We thank the breast-feeding mothers who participated in this studyand their families: other members ofthe field collection team, Karin MOstrom and Maureen Murtaugh: other laboratory staff. Marjorie BJackson and Sara Collins Couch: Constance M Capacchione and JeffreyR Backstrand for data management support: the hospital personnel who

were so cooperative and helpful in executing the study: and Lindsay H

Allen for review of the study design and thesis manuscripts.

References

2. Ferris AM. Dalidowitz DC. Ingardia CM. et al. Lactation outcome

in insulin-dependent diabetic women. I Am Diet Assoc 1988:88:3 17-22.

3. Saint L. Smith M. Hartmann PE. The yield and nutrient content

ofcolostrum and milk ofwomen from birth to I month post-partum.

BrI Nutr 1984:52:87-95.

4. Arthur PG. Smith M. Hartmann PE. Milk lactose, citrate, and glucoseas markers oflactogenesis in normal and diabetic women. I PediatrGastroenterol Nutr 1989:9:488-96.

5. Kulski 1K, Hartmann PE. Changes in human milk composition

during the initiation oflactation. Aust I Exp Biol Med Sci 1981:59:10 1-14.

6. Hartmann PE. Prosser CG. Physiological basis of longitudinalchanges in human milk yield and composition. Fed Proc 1984:43:2448-53.

7. Neville MC. Allen IC. Archer PC. et al. Studies in human lactation:milk volume and nutrient composition during weaning and lacto-

genesis. Am I Clin Nutr 1991:54:81-92.8. Bitman I. Hamosh M. Hamosh P. et al. Milk composition and vol-

ume during the onset of lactation in a diabetic mother. Am I Clin

Nutr I 989:50:1364-9.9. Frohlich A. Dzialoszynski LM. Insulin dependence ofmilk formation

and lactose biosynthesis. Clin Chim Acta 1973:47:33-8.

10. Nowak I, Dzialoszvnski LM. Effect ofexperimental alloxan diabeteson the secretion and composition of goat milk. Acta Physiol Pol

I967:18:488-97.

1 1. Sud SC. Effect of alloxan diabetes on the lactational performance

in rats. Indian I Exp Biol 1972:10:389-90.12. Butte NF. Garza C. Burr R. Goldman AS. Kennedy K. Kitzmiller

IL. Milk composition ofinsulin-dependent diabetic women. I Pediatr

Gastroenterol Nutr 1987:6:936-41.13. Stubblefield NE. Algert SI. Nutrition management ofdiabetes during

pregnancy. Top Clin Nutr 1988:3:58-63.14. Rosenfeld CR. Barton MD. Meschia G. Effects ofepinephrine on

distribution ofblood flow in the pregnant ewe. Am I Obstet Gynecol

1976:124:156-63.

15. Houvenaghel A. Action ofcatecholamines on blood flow through

the mammary gland in unanesthetized lactating small ruminants.Arch lnt Pharmacodvn Ther 1970:186:190-1.

16. Jones RG. Ilic V. Williamson DH. Regulation oflactating-rat mam-mary gland lipogenesis by insulin and glucagon in vivo. Biochem I

I984:223:345-5 I.

17. Linzell IL. Peaker M. Day-to-day variations in milk composition

in the goat and cow as a guide to the detection ofsubclinical mastitis.Br Vet I 1972:128:284-95.

18. Morriss FH. Physical-chemical changes in human milk during the

course oflactation. In: Howell RR, Morriss FH. Pickering LK. eds.Human milk in infant nutrition and health. Springfield. IL: CharlesC Thomas. 1986:51-65.

19. Ferns AM. Neubauer SH. Bendel RR. Green KW, Ingardia CC.

Reece EA. Perinatal lactation protocol and outcome in mothers withand without insulin-dependent diabetes mellitus. Am I Clin Nutr1993:57.

20. Gifford-lorgensen RA. Borchert I. Hassanein R. Tilzer L. Eaks GA.Moore WV. Comparison of five glucose meters for self-monitoring

ofblood glucose by diabetic patients. Diabetes Care 1986:9:70-6.21. Hare 1W, White P. Gestational diabetes and the White classification.

Diabetes Care 1980:3:394.22. Service Fl. What is “tight control” of diabetes? Goals. limitations

and evaluation oftherapy. Mayo Clin Proc 1986:61:792-5.23. Service Fl. Molnar GD. Rosevear 1W, Ackerman E. Gatewood LC.

Taylor WF. Mean amplitude ofglycemic excursions. a measure of

diabetic instability. Diabetes I 970:19:644-55.24. Molnar GD. Taylor WF. Langworthy A. On measuring the adequacy

ofdiabetes regulation: comparison ofcontinuously monitored bloodglucose patterns with values at selected time points. Diabetologia1974:10:139-43.

at HO

SP

ITA

L GR

AL U

NIV

ER

SIT

AR

IO A

LICA

NT

E B

IBLIO

TE

CA

on July 18, 2014ajcn.nutrition.org

Dow

nloaded from

60 NEUBAUER ET AL

25. Neville MC. Keller RP, Seacat I. Casey CE. Allen IC. Archer P.

Studies on human lactation. I. Within-feed and between breast vari-ation in selected components of human milk. Am I Clin Nutr

I984:40:635-46.

26. Hambraeus L. Forsum E. Abrahamsson L. L#{246}nnerdal,B. Automatic

total nitrogen analysis in nutritional evaluations using a block diges-tor. Anal Biochem 1976:72:78-85.

27. Linzell IL. Peaker M. Efficacy ofthe measurement ofthe electrical

conductivity ofmilk for the detection ofsubclinical mastitis in cows:

detection ofinfected cows at a single visit. Br Vet I 1975:131:447-

61.28. Hibberd CM, Brooke 0G. Carter ND. Haug M. Harzer G. Variation

in the composition ofbreast milk during the first 5 weeks of lactation:implications for the feeding of preterm infants. Arch Dis Child

1982:57:658-62.

29. SAS Institute Inc. SAS user’s guide: basics, version 5 edition. Cary.NC: SAS Institute Inc. 1985.

30. SAS Institute Inc. SAS user’s guide: statistics, version 5 edition. Cary.NC: SAS Institute Inc. 1985.

3 1. Winer BI. Statistical principles in experimental design. 2nd ed. New

York: McGraw-Hill, 1971.32. Cohen I. Statistical power analysis for the behavioral sciences. New

York: Academic Press, 1969.33. Hambraeus L. L#{246}nnerdalB. Forsum E. Gebre-Medhin M. Nitrogen

and protein components of human milk. Acta Paediatr Scand1978:67:561-5.

34. Healy DL. Rattigan S. Hartmann PE. Herington AC, Burger HG.

Prolactin in human milk: correlation with lactose. total protein. and

s-lactalbumin levels. Am I Physiol 1980:238:E83-6.35. Peaker M. Secretion of ions and water. In: Mepham T. ed. Bio-

chemistry oflactation. New York: Elsevier Science Publishers, 1983:285-305.

36. Procianoy RS. Fernandes-Filho PH. Lazaro L. Sartori NC. Factors

affecting breastfeeding: the influence of Caesarian Section. 1 TropPediatr 1984:30:39-42.

37. Kulski 1K. Smith M, Hartmann PE. Normal and caesarian section

delivery and the initiation of lactation in women. Aust I Exp BiolMed Sci 198 1:59:405-12.

38. Nommsen LA, Lovelady CA. Heinig Ml. L#{246}nnerdalB. Dewey KG.

Determinants of energy. protein. lipid. and lactose concentrationsin human milk during the first 12 mo of lactation: the DARLING

Study. Am I Clin Nutr 199 1:53:457-65.39. Neville MC. Keller. R, Seacat I. et al. Studies in human lactation:

milk volumes in lactating women during the onset of lactation and

full lactation. Am I Clin Nutr 1988:48:1375-86.40. Butte NF, Garza C. StufflE, Smith EO, Nichols BL. Effect of ma-

ternal diet and body composition on lactational performance. AmI Clin Nutr 1984:39:296-306.

41. DeCarvalho M, Robertson S. Friedman A. Klaus M. Effect of fre-

quent breast-feeding on early milk production and infant weightgain. Pediatrics 1983:72:307- 1 1.

42. Lammi-Keefe Cl. Ferris AM. Jensen RG. Changes in several com-ponents in human milk every four hours from 0600 to 2200. 1 PediatrGastroenterol Nutr 1989:1 1:83-8.

43. Dewey KG. L#{246}nnerdalB. Milk and nutrient intake of breast-fedinfants from 1 to 6 months: relation to growth and fatness. I PediatrGastroenterol Nutr I 983:2:497-506.

44. Ferns AM. Dotts MA. Clark RM, Ezrin M, Jensen RG. Macronu-

trients in human milk at 2. 12 and 16 weeks postpartum. I Am DietAssoc 1988:88:694-7.

45. Allen IC, Keller RP, Archer P. Neville MC. Studies in human lac-tation: milk composition and daily secretion rates of macronutrientsin the first year of lactation. Am I Clin Nutr 199 1:54:69-80.

46. Carneiro TA. Dutra de Oliveira JE. Nutritional studies in humanlactation in Brazil. 1. Chemical composition ofbreast milk. I Trop

Pediatr Environ Child Health 1973:19:384-7.47. L#{246}nnerdalB. Forsum E. Gebre-Medhin M, Hambraeus L. Breast

milk composition in Ethiopian and Swedish mothers. II. Lactose,nitrogen. and protein contents. Am I Clin Nutr 1976:29:1134-41.

48. L#{246}nnerdal B. Woodhouse LR. Glazier C. Compartmentalization

and quantitation of protein in human milk. I Nutr 1987:1 17:1385-95.

49. Ross SA. Clark RM. Nitrogen distribution in human milk from 2to 16 weeks postpartum. I Dairy Sci 1985:68:3199-201.

50. L#{246}nnerdalB. Forsum E. Hambraeus L. A longitudinal study of theprotein. nitrogen. and lactose contents ofhuman milk from Swedishwell-nourished mothers. Am I Clin Nutr 1976:29:1 127-33.

5 1 . Tolstoi E. The relationship ofthe blood glucose to the concentrationof lactose in the milk of lactating diabetic women. I Clin Invest

1935: 14:863-6.

52. Lemons IA. Moye L. Hall D. Simmons M. Differences in the com-

position of preterm and term human milk during early lactation.Pediatr Res 1982:16:113-7.

53. Tomarelli RM. Osmolality. osmolarity. and renal solute load of infant

formulas. I Pediatr 1976:88:454-6.54. Dorea 1G. Cabral MS. DaGloria M. Wright M. Rodrigues KH. Os-

molalities of bottle- and breast-milk fed to poor urban Brazilianinfants. Ann Trop Paediatr 1988:8:181-3.

55. Keenan BS. Buzek SW. Garza C. Potts E. Nichols BL. Diurnal andlongitudinal variations in human milk sodium and potassium: im-

plication for nutrition and physiology. Am I Clin Nutr 1982:35:527-34.

56. Zuppa AA. Tornesello A. Papacci P. et al. Relationship between

maternal parity. basal prolactin levels and neonatal breast milk in-take. BioI Neonate 1988:53:144-7.

57. Matheny R. Picciano MF. Feeding and growth characteristics ofhuman milk-fed infants. I Am Diet Assoc 1986:86:327-31.

at HO

SP

ITA

L GR

AL U

NIV

ER

SIT

AR

IO A

LICA

NT

E B

IBLIO

TE

CA

on July 18, 2014ajcn.nutrition.org

Dow

nloaded from